Objective:

To discuss the ethics of integrating AI into pathology and the management of AI validation data.

Approach:

• Expert Debate: Experts debate the balance between automation and human judgment in pathology, emphasizing the importance of trust between pathologists and AI.
• Case Study: The discussion includes examples from primary care where AI tools assist without replacing clinical judgment.
• Pathology Challenges: Pathologists face unique challenges due to their specialized knowledge, which complicates the integration of AI.
• Trust and Data: Trust in AI tools is linked to understanding their development and validation, requiring pathologists to ask critical questions about AI performance.

Key Findings:

• Trust between pathologists and AI will evolve over time as confidence in technology grows.
• The gold standard in pathology remains the judgment of experienced pathologists.
• Pathologists must determine which tasks can be delegated to AI while ensuring quality patient care.
• Cognitive biases of pathologists can influence their interpretation of AI recommendations.
• A formal due diligence framework is necessary for evaluating AI technologies in clinical use.

Interpretation:

The integration of AI in pathology requires careful consideration of human oversight, trust, and the unique challenges posed by the specialized nature of the field.

Limitations:

• The complexity of pathology expertise makes it difficult to translate knowledge into AI systems.
• Skepticism towards AI may shift to overconfidence as reliance on technology increases.

Conclusion:

The evolution of AI in pathology must be guided by pathologists' standards and the need for maintaining diagnostic quality.